I have an interesting puzzle for all you geeks.

Is there some unit of measurement that can be defined in absolute terms without reference to some other relatively arbitrary unit of measurement?

I'll use some counter examples. The definition of celsius (temperature) is based on the freezing point of water (0 C) and the boiling point (100 C). So, to exactly calibrate a thermometer, you could get some distilled water and observe the thermometer as you freeze the water. You could then heat the water to boiling. Viola! Your thermometer is calibrated.

Except... The boiling point of water varies by atmospheric pressure. The definition is defined "at sea level" - 29.92 inches of mercury. Ug. Inches? And under what gravitational pull?

Let's use another counter-example. You could define a unit of length based as the wavelength of a particular light. Okay, that's fine. Now you need an absolute way to generate the light in question. There are a number of objects that when heated radiate light. Great! We're almost there. Now, do you need the heated object at a specific temperature or pressure to produce the appropriate light?

So, a little homework: can you define a unit of measurement, either mass, length or time, that can be defined by some fixed method -- some method that would be true even under dramatically different conditions (different temperatures, different, atmospheric pressures, different gravitational forces, etc). Basically, a system that would work on Earth or on the dark side of the moon. Answers only work if they're practical. Defining your unit in terms of the mass of a hydrogen atom is worthless unless you can also demonstrate a reasonable means of counting hydrogen atoms.

Gary, SPL requires one Gary available to measure, and said Gary must be able to achieve a standard length under varying circumstances -- which I doubt happens. So please, SPLs will not survive my homework assignment.

Note that I haven't come up with any examples that work. I think the best chance is the wavelength of light produced by some light-emitting object, if there are objects that emit light at a fixed wavelength over a wide variety of temperatures and pressures.

There might be some means that make use of critical point and triple point. However, I question whether these methods would pass the practical use test given the extreme temperature and/or pressure ranges involved. That is, elements that have a relatively low critical vapor-liquid atmospheric value tend to have unreasonably low temperature values. For instance, hydrogen only requires 12.8 atmospheres but neg 240 degrees C.

Anyone have a good answer?